Split Thread: Electronic voting

[theprestige]

Relevant xkcd.

[LSSBB]

Originally Posted by theprestige

Relevant xkcd.

...already posted by another.

[jimbob]

Originally Posted by LSSBB

...already posted by another.

but worth reiterating

[Upchurch]

Originally Posted by theprestige

Relevant xkcd.

Originally Posted by LSSBB

...already posted by another.

Originally Posted by jimbob

but worth reiterating

Vote for Randall!

A Relevant XKCD on Every Page!

[theprestige]

Originally Posted by LSSBB

...already posted by another.

I searched for 'xkcd' and didn't find anything. Clearly this forum needs more blockchain.

[WilliamSeger]

Originally Posted by psionl0

Administrative weaknesses are a separate issue to the question of technical feasibility.


You are clearly posting about a completely different system of electronic voting.

Well, unless I missed it, you haven't yet described the system that you have in mind in sufficient detail to say if it avoids known generic problems.

[JoeMorgue]

Originally Posted by Upchurch

Vote for Randall!

A Relevant XKCD on Every Page!

If some sort of acopolyse wiped out everyone on Earth but the nerds, I swear after a few generations the spoken language with but nothing but "Damok and Jalad at Tenagra" style pop culture references and the written language would be pictographs based on xkcd and SMBC comics.

[Upchurch]

I have a friend at Defcon right now. The first talk he attended was about election hacking. Below is a picture he took of a voting machine along with his comment:

Quote:

Your voting machine security sucks...unless you like gifs!

[Leftus]

Originally Posted by Upchurch

And, just to head off the obvious counter argument, yes, you could do this with paper voting as well. The problem is, you need far more people to pull it off. If one guy keeps getting back into line to vote over and over, poll workers are probably going to pick up on it.

Online, all you need is one person with a botnet to do the thing.

Just to expand on this abit as well ...

Let's say they are going to allow e-voting registration to start tomorrow. The bad actors can register before you, most likely, as they don't need all that sleep stuff. So then to get your voting ID back you will likely have to provide all sorts of various ID documents to prove that you are who the government says you are. It could take months to straighten it all out.

The problem is not "will it all be added up correctly" which seems to be the focus of the blockchain. The problem is how do we determine that vote X was cast by an eligible person? Blockchain can not answer that.

[PhantomWolf]

Originally Posted by paulhutch

Very true.

However despite security failures being a continuously repeating problem in software systems for 4+ decades they are still a nearly daily occurrence. Of course then there's the backlash when a company improves security and the user base screams bloody murder because it's slightly less convenient. Frankly software engineers, myself included, are stuck between a rock and a hard place. Which is why although I agree on the over simplification of the comic strip I do believe strongly that electronic voting is still a very long way from being secure enough to rely on.

There is a guy I see advertising his security lectures on YT quite a bit, and one of the things he says is that "It is impossible to make a system that is totally secure against the bad guys, and totally insecure for the good guys."

He's right, you can't have a way in for the good guys and still expect to totally keep out the bad guys.

[psionl0]

Originally Posted by jimbob

If you have the resources, or are, say the people running the election, then a blockchain system is easy to set up to give the "desired" results with little chance of detection.

What a load of ****! The one part of the electronic system that can't be setup to give the desired results is the blockchain itself and if you read anything I posted then you would know why.

I know that electronic voting is not ready to be trusted in the real world. However, this desperate determination that everybody has to prove that absolutely . . . every . . . single . . . word that I have uttered on this subject to be wrong is insane!

[Upchurch]

Originally Posted by psionl0

I know that electronic voting is not ready to be trusted in the real world. However, this desperate determination that everybody has to prove that absolutely . . . every . . . single . . . word that I have uttered on this subject to be wrong is insane!

FWIW, I don’t think you are wrong about the workings of blockchain. I just think blockchain is largely irrelevant to the problems with online voting. Continuing to pitch it makes about as much sense as arguing about the font used in the voting device.

[jimbob]

Originally Posted by psionl0

What a load of ****! The one part of the electronic system that can't be setup to give the desired results is the blockchain itself and if you read anything I posted then you would know why.

I know that electronic voting is not ready to be trusted in the real world. However, this desperate determination that everybody has to prove that absolutely . . . every . . . single . . . word that I have uttered on this subject to be wrong is insane!

I will rephrase what I was meaning.


It is easy to set up the system *around* blockchain to give the desired results.

Which renders the integrity of the blockchain process in isolation rather moot

ETA Or what Upchurch said in the post above.

[psionl0]

Originally Posted by Upchurch

Continuing to pitch it makes about as much sense as arguing about the font used in the voting device.

Attempting to answer sensible questions about blockchain is not "continuing to pitch it".

[jimbob]

Originally Posted by psionl0

Attempting to answer sensible questions about blockchain is not "continuing to pitch it".

It is, as blockchain isn't the obvious problem - just what would need to be done to use it in a voting system

[psionl0]

Originally Posted by jimbob

It is, . . .

I should put anybody who asks question or posts incorrect information about the blockchain on ignore?

[Lplus]

Originally Posted by JoeMorgue

I wish we'd get over our fear of online voting. 99% of America has a web capable device near literally grafted to them every waking moment of their life and even for the small percentage that don't that just means nearly every home, office, library, and internet cafe becomes a voting booth.

Voting website opens a 8 a.m. Eastern Standard time on a Monday and closes at 8 p.m Eastern Standard time the following Saturday. You login to the website and it gives you breakdown all your eligible elections. You pick the candidates, hit a big "Vote" button. Sign out. That's over 5 full days where the only requirement is "Get to some kind of online capable device for maybe 5 minutes" Over night on Saturday all the votes can be tallied, on Sunday verified by some external auditing system to give people the warm and fuzzy, and the results announced Sunday night.

Ok, I read all the stuff above, but no one has asked the question - Why do you feel the existing system should be changed?

Eh, maybe that question was posed in the thread this was split from

[jimbob]

Originally Posted by psionl0

I should put anybody who asks question or posts incorrect information about the blockchain on ignore?

OK, a different way to get my point home.

How does blockchain prevent a corrupt voting register creating phantom registered voters that sit in the system and then verifying them with blockchain? The blockchain itself can be perfectly reliable, but it can't know that it's been fed bad starting data.

[jimbob]

Originally Posted by psionl0

I should put anybody who asks question or posts incorrect information about the blockchain on ignore?

OK, a different way to get my point home.

How does blockchain prevent a corrupt voting register creating phantom registered voters that sit in the system and then verifying them with blockchain? The blockchain itself can be perfectly reliable, but it can't know that it's been fed bad starting data.

[Upchurch]

Originally Posted by psionl0

Attempting to answer sensible questions about blockchain is not "continuing to pitch it".

You cherrypick the questions you attempt to answer about it, favoring the questions that allow you to promote its use and ignoring questions about the limitations of its usefulness.

[psionl0]

Originally Posted by Upchurch

You cherrypick the questions you attempt to answer about it, favoring the questions that allow you to promote its use and ignoring questions about the limitations of its usefulness.

I am under no obligation to respond to strawman arguments.

In your case it wouldn't make any difference. I could say "you are absolutely right about everything" and you would still find fault with what I say.

[WilliamSeger]

Originally Posted by psionl0

I am under no obligation to respond to strawman arguments.

In your case it wouldn't make any difference. I could say "you are absolutely right about everything" and you would still find fault with what I say.

Well, if all you want to talk about is blockchain, rather than electronic voting , here's something that's bothered me about it since I first heard about it: The blockchain security and the "mining" are based on the proof-of-work mathematical puzzle you mentioned being very hard because the only known way to solve it is trial and error. It's that "only known way" that bothers me when it comes to math. What if someone finds a shortcut or just some pattern in existing blocks that allows them to solve the puzzle much faster?

[psionl0]

Originally Posted by WilliamSeger

Well, if all you want to talk about is blockchain, rather than electronic voting , . . .

Not true. I have also discussed electronic voting tickets and how they can be stolen, misallocated or filled out by malware (presumably while displaying false information to the voter). I have also discussed the need for a process in case a voter complains that the vote on the blockchain isn't the way they voted.

Of course, if you only read what others say I have posted then you would never have believed that.

Originally Posted by WilliamSeger

here's something that's bothered me about it since I first heard about it: The blockchain security and the "mining" are based on the proof-of-work mathematical puzzle you mentioned being very hard because the only known way to solve it is trial and error. It's that "only known way" that bothers me when it comes to math. What if someone finds a shortcut or just some pattern in existing blocks that allows them to solve the puzzle much faster?

The key to blockchain security (in bitcoin) is the SHA256 algorithm. It takes a stream of characters as its input and outputs a 256 bit "hash". If just one bit of the input stream is altered then a completely different hash is produced. So far, there is no known way to reverse the hashing algorithm. That is, given a hash code, there is no known way to deduce the original character stream. So you can only keep guessing until you get the answer "close enough". If a mathematical method were ever discovered to reverse hash codes then the viability of blockchains would be the least of our problems.

Note that there is no backdoor method of bypassing hash codes. Being open source software, any attempt to include a back door would have been discovered almost immediately.

[Upchurch]

Originally Posted by psionl0

I am under no obligation to respond to strawman arguments.

What strawman arguments?

Originally Posted by psionl0

In your case it wouldn't make any difference. I could say "you are absolutely right about everything" and you would still find fault with what I say.

Whoa. Ad hominem.

I’m not finding any fault with your descriptions or understanding of block chain or how it could be used for electronic voting. What i, and others, have pointed out is that it is largely immaterial to the actual problems involved, to which you have already agreed.

The difference seems to be that you think it will ever make sense to have a reliable electronic voting system, if only some of the technology were better. I’m telling you that will never happen so long as humans have any access to the system at any level*. Give that there is a profit to be made in rigging elections only assures that even more.



* Hell, you don’t even need access to the voting system, if you have access to the voters and a convincing spoof.

[jimbob]

As per Upchurch, I can accept that the blockchain algorithm is mathematically perfect.

I can accept for the sake of argument that a system would enable those people who are issued with votes to verify that their vote was cast as they intended whilst simultaneously allowing other people to check that only people on the electoral roll have voted.

The problem is that this still leaves massive loopholes.

How do you preserve anonymity if you want to check that all the votes for a particular candidate are counted at one vote per person?

How do you ensure that all the voters exist outside of the electoral roll?

How do you protect against spear phishing?

Also, open source software still has security issues. At work, we have had Linux security patches issued.

[WilliamSeger]

Originally Posted by psionl0

Not true. I have also discussed electronic voting tickets and how they can be stolen, misallocated or filled out by malware (presumably while displaying false information to the voter). I have also discussed the need for a process in case a voter complains that the vote on the blockchain isn't the way they voted.

Of course, if you only read what others say I have posted then you would never have believed that.


The key to blockchain security (in bitcoin) is the SHA256 algorithm. It takes a stream of characters as its input and outputs a 256 bit "hash". If just one bit of the input stream is altered then a completely different hash is produced. So far, there is no known way to reverse the hashing algorithm. That is, given a hash code, there is no known way to deduce the original character stream. So you can only keep guessing until you get the answer "close enough". If a mathematical method were ever discovered to reverse hash codes then the viability of blockchains would be the least of our problems.

Note that there is no backdoor method of bypassing hash codes. Being open source software, any attempt to include a back door would have been discovered almost immediately.

It wouldn't necessarily need to reverse a given hash. Reducing the number of candidate nonces to be tested would be enough to destroy the security of the blockchain if it allowed a bad actor to create an altered chain of blocks faster than the rest of the network.

[jimbob]

could you say that in English, please?

Although that does highlight the problem of transparency - many experts are highly skeptical of the idea, but non experts are supposed to take the assurances of the enthusiasts on trust ( most people don't understand blockchain).

[Leftus]

https://www.msn.com/en-us/news/polit...cid=spartanntp

670 votes cast. 276 Registered voters.

[Blue Mountain]

Originally Posted by Leftus

https://www.msn.com/en-us/news/polit...cid=spartanntp

670 votes cast. 276 Registered voters.

It was later discovered the voting machine had the wrong number of voters. The true number of eligible voters was much higher than 670.

[psionl0]

Originally Posted by WilliamSeger

It wouldn't necessarily need to reverse a given hash. Reducing the number of candidate nonces to be tested would be enough to destroy the security of the blockchain if it allowed a bad actor to create an altered chain of blocks faster than the rest of the network.

If you are referring to 51% attacks then that has already been discussed. http://www.internationalskeptics.com...6#post12384356

Originally Posted by Leftus

https://www.msn.com/en-us/news/polit...cid=spartanntp

670 votes cast. 276 Registered voters.

I think it has been pretty well established that the existing electronic voting machines were set up by criminals who's main objectives were hackability and opaqueness.

[WilliamSeger]

Originally Posted by psionl0

If you are referring to 51% attacks then that has already been discussed. http://www.internationalskeptics.com...6#post12384356

No, I'm referring to the fact that it is theoretically possible to alter a block if you alter all the blocks that follow it, which requires finding new nonces for each, and then add another block that makes it longer than any chain that any node already has. That's because when a node sees a difference in chains distributed from other nodes, the longest (most work) chain is taken as the valid one. The only thing that prevents that is the (current) inability to do that faster than all the other nodes combined can add new blocks to the chain, i.e. you'd need more computing power than all the rest of the nodes to get ahead of them. If someone found some sort of filter that reduced the number of trials required by a couple orders of magnitude, then it might be possible.

[psionl0]

Originally Posted by WilliamSeger

No, I'm referring to the fact that it is theoretically possible to alter a block if you alter all the blocks that follow it, which requires finding new nonces for each, and then add another block that makes it longer than any chain that any node already has. That's because when a node sees a difference in chains distributed from other nodes, the longest (most work) chain is taken as the valid one. The only thing that prevents that is the (current) inability to do that faster than all the other nodes combined can add new blocks to the chain, i.e. you'd need more computing power than all the rest of the nodes to get ahead of them.

I think you have answered your own objection. There is not enough computing power to alter more than the most recent block or two and even then, you would have to be able to compute nonces at least twice as fast as the rest of the network combined. Six confirmations (the sixth most recent block) is usually regarded as the the point where we can say that alteration becomes "impossible". Even if you could get at the block, you would not be able to alter the votes in the block because each vote is digitally signed. The most you could do is delete a vote from the block but then other nodes would pick it up and add it to the blockchain again.

Originally Posted by WilliamSeger

If someone found some sort of filter that reduced the number of trials required by a couple orders of magnitude, then it might be possible.

Aaaand . . . we are back to cracking the SHA 256 hashing algorithm again. Trust me, if that ever happens then the whole world will know about it.

[jimbob]

Originally Posted by psionl0

I think you have answered your own objection. There is not enough computing power to alter more than the most recent block or two and even then, you would have to be able to compute nonces at least twice as fast as the rest of the network combined. Six confirmations (the sixth most recent block) is usually regarded as the the point where we can say that alteration becomes "impossible". Even if you could get at the block, you would not be able to alter the votes in the block because each vote is digitally signed. The most you could do is delete a vote from the block but then other nodes would pick it up and add it to the blockchain again.


Aaaand . . . we are back to cracking the SHA 256 hashing algorithm again. Trust me, if that ever happens then the whole world will know about it.

What happens if you start at the first vote... because you have compromised the system in a way that gives you a head start of a few seconds?

Regardless of this, it seems a bit like saying that the Yamato would have been able to defeat any USN battleship. Arguable, but irrelevant if the attack vector is something completely different.

[WilliamSeger]

Originally Posted by psionl0

Aaaand . . . we are back to cracking the SHA 256 hashing algorithm again. Trust me, if that ever happens then the whole world will know about it.

You don't understand what I mean, which is not "cracking" SHA256. I'm simply talking about a filter that would reduce the average number of trial nonces that would need to be tested to find one that produced the required number of leading zeros in the hash. For all I know, some sort of neural net might find a usable pattern in past blocks. I'm not saying that I expect that to happen; I'm saying that's one of the reasons I'm not buying any bitcoins -- rather far down the list, actually.

[PhantomWolf]

Originally Posted by psionl0

I think it has been pretty well established that the existing electronic voting machines were set up by criminals who's main objectives were hackability and opaqueness.

Let's look at some of the recent criminal activity with Elections. In the 2016 Election, the Russians stole identity data from thousands of USers via FB and Cambridge Analytics. We also know that they were able to hack a number of State's voter registration systems.

In your system, how would you prevent them from doing the same thing, determine which voters are unlikely to vote or register to vote, then hack into the system and create registered voters for those people, which they could then use to create votes in the election?

[psionl0]

Originally Posted by PhantomWolf

Let's look at some of the recent criminal activity with Elections. In the 2016 Election, the Russians stole identity data from thousands of USers via FB and Cambridge Analytics. We also know that they were able to hack a number of State's voter registration systems.

In your system, how would you prevent them from doing the same thing, determine which voters are unlikely to vote or register to vote, then hack into the system and create registered voters for those people, which they could then use to create votes in the election?

I don't have a satisfactory answer to this question. Online Voter registration and identity theft are not issues that can be solved by a blockchain (AFAIK).

i suspect that if done properly, fraudulent registration could be minimized but that doesn't seem to be one of the objectives of those who set these systems up.

[Upchurch]

Originally Posted by PhantomWolf

In your system, how would you prevent them from doing the same thing, determine which voters are unlikely to vote or register to vote, then hack into the system and create registered voters for those people, which they could then use to create votes in the election?

And then, once you’re done with that, you then have to do what electronic security experts do:

1. Ask yourself: How would I beat this system?
2. Fix that.
3. Goto 1.

No exit statement. If you ever hit “I can’t” to 1, it’s time to bring in new people because you’re missing something.

[jimbob]

Originally Posted by PhantomWolf

Let's look at some of the recent criminal activity with Elections. In the 2016 Election, the Russians stole identity data from thousands of USers via FB and Cambridge Analytics. We also know that they were able to hack a number of State's voter registration systems.

In your system, how would you prevent them from doing the same thing, determine which voters are unlikely to vote or register to vote, then hack into the system and create registered voters for those people, which they could then use to create votes in the election?

Yes, that was my first thought, but then I realised that there is also nothing stopping the electoral register having phantom voters and just assigning them votes as desired.

Depending on where the system is compromised.

[Upchurch]

Originally Posted by psionl0

I don't have a satisfactory answer to this question.

That's because there are none.

I feel like this needs another watch by the thread at large:

YouTube Video This video is not hosted by the ISF. The ISF can not be held responsible for the suitability or legality of this material. By clicking the link below you agree to view content from an external website.

I AGREE

[psionl0]

Originally Posted by Upchurch

That's because there are none.

An absolutely positively certain fact. Why? Because you say so.

Originally Posted by Upchurch

I feel like this needs another watch by the thread at large:

What does a repeat screening of an out of date video prove? Does repetition make it righter?